Simon Kuznets, from "The Measurement of Economic Growth," 1947

II

Growth is a concept whose proper domicile is the study of organic units, and the use of the concept in economics is an example of that prevalent employment of analogy the danger of which have been so eloquently stressed recently by Sidney Hook. Nevertheless it might be useful to see how the concept is defined in the field of its original habitat as a clue to what it may mean in application to human societies. 

To use one definition, growth signifies "a process, indirectly resulting from chemical, osmotic and other forces, by which material is introduced into the organism and transferred from one part of it to another." By analogy, economic growth is a process by which economic material is introduced into a nation's economy and transferred from one part of it to another. 

If long-standing statistical practices are any indication, the economic material in question is most directly represented by what economists designate productive resources: natural, irreproducible goods, such as land, mineral deposits, rivers, and waterways; population; and reproducible wealth, in the form of all types of equipment, inventories, and so forth, including (from the standpoint of a given nation) effective economic claims upon other nations. Just as the growth of an organism would be measured by the increase of its weight, height, number of cells, and so forth, so the growth of a nation would be gauged by additions to its wealth and population. 

That sustained increases, in natural resources viewed as materials and means of production, in population viewed as labor supply, and in reproducible resources viewed as accumulated capital, are, each separately and all together, indications of a nation's economic growth can hardly be disputed. But we are interested in reliable measures, indexes that would state unequivocally not only the existence of economic growth over a certain period but also its rate for comparison with other periods or among nations. Identification of economic growth with increase in the stock of resources, when examined from this standpoint, suffers from several major shortcomings. 

The first is the inherent difficulty of measurement—particularly true of natural resources. The controversies among geologists concerning the definition and amounts of proven and probable resources and the apparent difficulties in measuring the changing quality of the soil in agriculture are significant not because of transient disagreements about magnitudes involved. They are important because in the case of resources that have not been brought into active economic circulation, secure knowledge of magnitudes may be impossible—in the sense that society is not forced to value them and may be unwilling to incur the costs of establishing their magnitudes unequivocally in any other way. 

The second, more important, difficulty is that some productive resources, that is, some factors that can be viewed as contributing to economic production, are by their very nature not measurable. The most important productive resource available in modern society is the stock of technological knowledge embodied in tangible records and in the personal skills and habits of the population. One might argue that this is not a separate resource but rather part and parcel of population as a productive resource or of the natural deposits and of accumulated reproducible capital. This does not remove the difficulty that this separate resource or aspect of other resources is not measurable—that, in other words, the simple natural units, such as numbers of people, tons of minerals, acres of land, horsepower of machinery, and so forth, in which we can measure amounts of productive resources in the categories suggested fail to reflect what is perhaps the most important item in the stock of economic resources. 

The third difficulty is that of finding a common base for combining into a whole the measures of separate categories, wide or narrow, of stocks of resources. Unless one is willing to claim that a single category is either determinative or symptomatic of all other quantitative aspects of economic growth, some way of combining the measures of the separate categories must be found. This is particularly important for modern economic societies in which historical experience suggests that movements of population, reproducible capital, and natural resources can diverge in rate, if not in direction, for the same nation over considerable periods. Yet how can one combine population numbers with dollar values of accumulated reproducible capital, or with B.T.U. equivalents of fuel deposits? 

It is important to note the source of this difficulty: that only a small proportion of the existing stock of even tangible resources enters, in its own form, into economic circulation. Not the resources themselves but their current services flow and are appraised in the process of economic circulation. For this reason, even for such categories of resources as carry economic magnitudes, for example, reproducible-wealth items (machinery, inventories, buildings, and so on), it is not easy to find a reliable appraisal of economic value: only a small portion of the total stock passes through current markets, and even that portion may be qualitatively different from the total. Anyone studying the successive censuses of wealth in the United States cannot but be impressed by the difficulty of finding a valuation basis roughly meaningful at any given time and consistent from time to time -—even for the given, narrowly circumscribed stock of economic resources included (exclusive of population, foreign debt, and, in most cases, of direct measures of inventories). 

Measures of the stock of resources, in so far as they can be secured, can be extremely useful as rough approximations to at least some of the determinants of economic growth of a nation, and, for lack of better measures, can often be used as symptoms of the existence or absence of economic growth. But the shortcomings just noted render them rather inefficient measures of the rate of growth. We turn now to search for more efficient measures, not on the plane of accumulated tangible resources in their natural units, but on the plane of economic production and circulation. This shift is necessary because no stable relations can be found (or at least, as far as I know, have been found) between sustained movements in the measurable stock of resources and in the magnitude of the total performance of national economies. 

III 

On this plane one begins by viewing the nation primarily as a production unit and assuming that the approach by way of production will yield the most comprehensive measure of the functioning of a nation's economy. Exchange, distribution, consumption, and accumulation can then be seen as stages in the circulation of economic goods whose total magnitudes are most likely, especially in the long run of economic growth, to be less comprehensive than the magnitude of total production. By production we mean the output of all scarce goods. Measurement of total production, as practiced in the old and rapidly growing literature on national income and product, can be, and has been, attempted at different stages of economic circulation: at the point of origin of goods in the producing units of the economy; at the point of payments flowing from producing units to the currently engaged productive factors; at the point of flow of these goods into ultimate consumption or capital accumulation. We need not consider here the problems in the different forms in which they emerge in each of these several approaches. It is sufficient to indicate their nature by that approach in which they are perhaps most clearly revealed—in the one in which we view total output as a sum of products flowing to the individuals and families who are the nation's ultimate consumers, and into net capital accumulation of various types, including additions to claims against foreign countries. 

Of the three major problems encountered in defining total output the first, that of scope, arises because production is maintained under the auspices and governance of several economic institutions, which are distinguished by differences in the whole complex of motives, rules, and measures that govern choices. Thus in our modern society, as in many societies of the past, at least three major institutions are to be distinguished: the family, the business enterprise, and the state. Unless a measure of total output is to reflect the growth of a given institution alone, it obviously should include all economic production within the family, the business enterprise, and the state. Yet most measures of national income note only market-bound output, including almost all state production but omitting large portions of productive activity which, not being market-bound and forming an integral part of family life, are not considered properly economic. There is a definite choice here between totals more comprehensive but less homogeneous and those less comprehensive but more homogeneous. 

However unimportant this difficulty may appear for short-term studies, in the long periods implied in measuring economic growth the problem is of too large a magnitude to be dismissed easily. Such long periods are characterized by important shifts in the weights of these different institutions, and reducing the scope of measurement will necessarily produce a substantial bias. Of the quantitatively impressive growth of total output in this country, as measured in the ordinary estimates of national income, a large part is to be associated with the extension of the business at the expense of the family sector. Consequently, one important prerequisite for a more efficient measurement of economic growth lies in the inclusion of such sectors of production that easily escape the statistical eye. As specific examples we may cite the capital formation involved in the work of American farmers in bringing virgin land into cultivation, or the work within the old- fashioned large family, so much of which has been taken over in recent decades by business firms. 

The second problem is that of obtaining an unduplicated total of all output. This problem might seem to have been solved in the definition of total output used here: as the sum of products flowing into ultimate consumption plus net additions to the stock of goods within the country and to claims against foreign countries. Yet this simple definition hides grave problems. 

The first emerges when we ask why flow of goods into ultimate consumption is considered in and of itself an unduplicated total. The reason presumably is that we view ultimate consumers, individuals and families, not as productive resources and machines but as human beings for the satisfaction of whose wants the economy operates. Hence any goods which they purchase cannot be viewed as being consumed in the productive process of turning out other goods, and hence for this total such duplication, as that between say the value of pig iron produced and the value of the bridge constructed with this iron, is impossible. 

But if this is the case, two parts of flow of goods to consumers are to be viewed with suspicion. One is that purchased by individuals who want it not as ultimate consumers but as producers. This may range all the way from such easily classifiable items as work clothing or transportation to the place of employment to such perplexing items as a luxury house or car considered indispensable to the functioning of its consumer as a producer (say as a business executive of high standing). A substantial part of what is currently defined as ultimate consumption is perhaps not final product but actually means of production. Its inclusion represents duplication, and quite likely the relative magnitude of such duplication is much greater in modern urban times than it was in the simpler economic civilizations of preindustrial societies. 

The second questionable sector of flow of goods to final consumers is the product of government. Unless we conceive government output as final by definition, for which there is no reason except convenience, it may be viewed either as services to ultimate consumers and hence a finished product, or as services to business or to society at large and hence an intermediate product which is in turn consumed in the production process and should not be included in a net total of economic output. But what are these services to ultimate consumers? Should we include, as I think we should, only services that are relevant to individuals as ultimate consumers and have a clear counterpart on private markets (for example, medical service, education, and the like)? Or should we include some parts of the general activity of the state in assuring internal peace and external integrity such as judiciary, police, military, and so forth? In the latter case, increased production of munitions by the state would be allowed to swell the unduplicated total of the net product of a nation's economy. Were we to include these types of services under the flow of goods to ultimate consumers, I would argue that intermediate products are counted in, that is, products that are used to provide the basis for further production rather than to serve the satisfaction of the ultimate consumer as such. Anyone who has looked at the astronomical figures of war expenditures in recent years or made reasonable forecasts of their magnitude for the future can easily see that the question raised involves not minutiae but large segments of current production, and that the way it is answered will affect materially the picture of economic growth in recent years. 

There are questions of somewhat different character and magnitude concerning grossness and netness in the second sector of total production as we define it, namely capital accumulation. The concept of net additions to stock of capital is quite clear. In practice, there is only gross output; and we have only the vaguest idea of the consumption of existing capital that should be used as an offset to derive net additions. The reason for this vagueness is exactly that which explained the difficulties in using stocks of resources as indexes of economic growth. Consumption of capital is a hidden process which is known only post factum and not too clearly even then. All we see and all that circulates is gross product. How much of reproducible capital, or particularly of some of the natural resources, has been consumed in the process is not visible, and many of the available measures are mere conventions. What is worse, for some types of capital no measures of consumption are at all available—as is the case with many natural resources either in public hands or in the hands of individual entrepreneurs not accustomed to proper accounting. It is more than likely that whereas current consumption of reproducible capital is often exaggerated in available data, that of nonreproducible capital is often underestimated. And the destructive effects of the intensive type of production characterizing industrial societies is not often fully reflected in the long-term estimates of what is presumably a net volume of total output. 

The third major problem in defining a measurable total of a nation's output is the valuation of its parts on a common basis. As distinct from the existing stock of resources, an overwhelming proportion of current output does pass through the market place and is assigned economic values. Nevertheless, difficulties arise. First, with comprehensive scope, the inclusion of that part of production that never appears on the market must be subjected to valuation comparable to that applied to market-bound goods—a difficulty that is resolved, if only approximately, by assigning to nonmarket goods the prices of analogous items that do appear on the market. Second, markets differ in the freedom with which prices are fixed in them and the range of difference is from purely competitive situations to the type of exclusive monopoly practiced by governments. Here the solution is much less easily found, for by the very nature of the case analogues cannot be so easily established. Finally, in so far as we speak of volumes relating to different periods, not only changing price levels but also shifting relative prices of various categories of goods are to be dealt with. The available techniques used in compiling index numbers, even disregarding questions of availability of price data, all encounter the difficulty that using the weights of one period as a base will yield results differing from those obtained by using weights of another period as a base. The consequence is that when, from one time point to another, shifts in the weight and composition of the aggregate of production have taken place, the rate of change between two points of time can be established only within certain limits—the limits indicated by using as base first the weights of the initial time point and then the weights of the terminal time point. Since economic growth, as repeatedly stressed, involves long time periods and since these periods are necessarily characterized by significant shifts in the composition of aggregate production and in the rela tive weights of various categories of goods in them, this limitation upon the determinacy of the measured rate is important. 

The three groups of problems noted all stem from a single source: conflict between the need for a single, consistent measure that would permit proper comparisons of magnitude or rate of growth for a national economy and the lack of measurability of directly observable economic reality. It is because family and household-bound activities take place away from the yardstick of the market, because the market mechanisms record the flow of goods not once but several times during a year and record also items that by no stretch of the imagination can be classified as goods, because the weights attached by markets to identical goods vary widely from time to time and place to place—that problems of scope, duplication, and valuation arise. Indeed, these problems may at first appear so grave as perhaps to make us think that we are trying to measure the immeasurable. Yet the questions posed can be answered, at least tentatively; and they are answered in terms of the basic types of uses to which the measures are to be put. 

IV 

A number that represents the total magnitude of a nation as an economic unit, whether for a given year or for a period long enough to permit observation of economic growth, is a raw datum the efficiency of which must be judged in terms of uses to which it is to be put, scientific purposes which it is to serve. ...